Benchmark search

Try these searches

Found 20 run configurations

Each row is a run configuration (model + hardware + framework + quantization); repeated runs of the same setup are merged — raw counts appear in each result’s source line.

ModelMiniMax-M3

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 5080 SUPER · 24GB · SGLang

Other

Today

4 tok/s

Decode

30 tok/s

Prefill

3.33 s

TTFT

24 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 5080 SUPER · 24GB · SGLang

Other

Today

5.3 tok/s

Decode

39.8 tok/s

Prefill

2.51 s

TTFT

24 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 5080 SUPER · 24GB · vLLM

Other

Today

4.1 tok/s

Decode

30.7 tok/s

Prefill

3.26 s

TTFT

24 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 5080 SUPER · 24GB · vLLM

Other

Today

5.5 tok/s

Decode

41.3 tok/s

Prefill

2.42 s

TTFT

24 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 5070 SUPER · 18GB · SGLang

Other

Today

3.2 tok/s

Decode

24 tok/s

Prefill

4.17 s

TTFT

18 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 5070 SUPER · 18GB · SGLang

Other

Today

4.2 tok/s

Decode

31.5 tok/s

Prefill

3.17 s

TTFT

18 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 5070 SUPER · 18GB · vLLM

Other

Today

3.3 tok/s

Decode

24.8 tok/s

Prefill

4.03 s

TTFT

18 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 5070 SUPER · 18GB · vLLM

Other

Today

4.4 tok/s

Decode

33 tok/s

Prefill

3.03 s

TTFT

18 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 5070 Ti · 16GB · SGLang

Other

Today

3.2 tok/s

Decode

24 tok/s

Prefill

4.17 s

TTFT

16 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 5070 Ti · 16GB · SGLang

Other

Today

4.2 tok/s

Decode

31.5 tok/s

Prefill

3.17 s

TTFT

16 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 5070 Ti · 16GB · vLLM

Other

Today

3.3 tok/s

Decode

24.8 tok/s

Prefill

4.03 s

TTFT

16 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 5070 Ti · 16GB · vLLM

Other

Today

4.4 tok/s

Decode

33 tok/s

Prefill

3.03 s

TTFT

16 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 5090 · 32GB · SGLang

Other

Today

6.4 tok/s

Decode

48 tok/s

Prefill

2.08 s

TTFT

32 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 5090 · 32GB · SGLang

Other

Today

8.5 tok/s

Decode

63.8 tok/s

Prefill

1.57 s

TTFT

32 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 5090 · 32GB · vLLM

Other

Today

6.6 tok/s

Decode

49.5 tok/s

Prefill

2.02 s

TTFT

32 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 5090 · 32GB · vLLM

Other

Today

8.8 tok/s

Decode

66 tok/s

Prefill

1.52 s

TTFT

32 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 4090 · 24GB · SGLang

Other

Today

3.6 tok/s

Decode

27 tok/s

Prefill

3.7 s

TTFT

24 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 4090 · 24GB · SGLang

Other

Today

4.8 tok/s

Decode

36 tok/s

Prefill

2.78 s

TTFT

24 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S SGLang 0.4.4 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ4_XS

L0 Self-reported

NVIDIA RTX 4090 · 24GB · vLLM

Other

Today

3.7 tok/s

Decode

27.8 tok/s

Prefill

3.6 s

TTFT

24 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ4_XS vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

MiniMax-M3 · IQ3_S

L0 Self-reported

NVIDIA RTX 4090 · 24GB · vLLM

Other

Today

5 tok/s

Decode

37.5 tok/s

Prefill

2.67 s

TTFT

24 GB

VRAM

Config (latest run)
Linux CUDA 12.6 IQ3_S vLLM 0.7.3 Flash Attention

Sources: Other · 1 runs · 1 independent sources

20 results

Related models
Related hardware
Related frameworks
Related quantizations